Detecting Gas Vapor Leaks Using Uncalibrated Sensors

Diaa Badawi, Tuba Ayhan, Sule Ozev, Chengmo Yang, Alex Orailoglu, Ahmet Enis Cetin

Research output: Contribution to journalArticle

Abstract

Chemical and infra-red sensors generate distinct responses under similar conditions because of sensor drift, noise or resolution errors. In this paper, we develop novel machine learning methods for detecting and identifying VOC and Ammonia vapor from time-series data obtained by uncalibrated chemical and infrared sensors. We process time-series sensor signals using deep neural networks (DNN). Three neural network algorithms are utilized for this purpose. Additive neural networks (termed AddNet) are based on a multiplication-devoid operator and consequently exhibit energy efficiency compared to regular neural networks. The second algorithm uses generative adversarial neural networks so as to expose the classifying neural network to more realistic data points in order to help the classifier network to deliver improved generalization. Finally, we use conventional convolutional neural networks as a baseline method. Our findings indicate that using raw time-series data obtained from uncalibrated sensors and processing them using deep-learning-based methods yield better results than using hand-crafted feature parameters.

Original languageEnglish (US)
Article number8883150
Pages (from-to)155701-155710
Number of pages10
JournalIEEE Access
Volume7
DOIs
StatePublished - Jan 1 2019

Fingerprint

Gases
Vapors
Neural networks
Sensors
Time series
Infrared radiation
Volatile organic compounds
Ammonia
Energy efficiency
Learning systems
Classifiers
Processing

Keywords

  • additive
  • and generative adversarial (GAN) neural networks
  • convolutional
  • sensor drift
  • time-series data analysis
  • VOC gas leak detection

ASJC Scopus subject areas

  • Computer Science(all)
  • Materials Science(all)
  • Engineering(all)

Cite this

Badawi, D., Ayhan, T., Ozev, S., Yang, C., Orailoglu, A., & Cetin, A. E. (2019). Detecting Gas Vapor Leaks Using Uncalibrated Sensors. IEEE Access, 7, 155701-155710. [8883150]. https://doi.org/10.1109/ACCESS.2019.2949740

Detecting Gas Vapor Leaks Using Uncalibrated Sensors. / Badawi, Diaa; Ayhan, Tuba; Ozev, Sule; Yang, Chengmo; Orailoglu, Alex; Cetin, Ahmet Enis.

In: IEEE Access, Vol. 7, 8883150, 01.01.2019, p. 155701-155710.

Research output: Contribution to journalArticle

Badawi, D, Ayhan, T, Ozev, S, Yang, C, Orailoglu, A & Cetin, AE 2019, 'Detecting Gas Vapor Leaks Using Uncalibrated Sensors', IEEE Access, vol. 7, 8883150, pp. 155701-155710. https://doi.org/10.1109/ACCESS.2019.2949740
Badawi D, Ayhan T, Ozev S, Yang C, Orailoglu A, Cetin AE. Detecting Gas Vapor Leaks Using Uncalibrated Sensors. IEEE Access. 2019 Jan 1;7:155701-155710. 8883150. https://doi.org/10.1109/ACCESS.2019.2949740
Badawi, Diaa ; Ayhan, Tuba ; Ozev, Sule ; Yang, Chengmo ; Orailoglu, Alex ; Cetin, Ahmet Enis. / Detecting Gas Vapor Leaks Using Uncalibrated Sensors. In: IEEE Access. 2019 ; Vol. 7. pp. 155701-155710.
@article{ab7e749bf2ef4a519f560240239d7f1f,
title = "Detecting Gas Vapor Leaks Using Uncalibrated Sensors",
abstract = "Chemical and infra-red sensors generate distinct responses under similar conditions because of sensor drift, noise or resolution errors. In this paper, we develop novel machine learning methods for detecting and identifying VOC and Ammonia vapor from time-series data obtained by uncalibrated chemical and infrared sensors. We process time-series sensor signals using deep neural networks (DNN). Three neural network algorithms are utilized for this purpose. Additive neural networks (termed AddNet) are based on a multiplication-devoid operator and consequently exhibit energy efficiency compared to regular neural networks. The second algorithm uses generative adversarial neural networks so as to expose the classifying neural network to more realistic data points in order to help the classifier network to deliver improved generalization. Finally, we use conventional convolutional neural networks as a baseline method. Our findings indicate that using raw time-series data obtained from uncalibrated sensors and processing them using deep-learning-based methods yield better results than using hand-crafted feature parameters.",
keywords = "additive, and generative adversarial (GAN) neural networks, convolutional, sensor drift, time-series data analysis, VOC gas leak detection",
author = "Diaa Badawi and Tuba Ayhan and Sule Ozev and Chengmo Yang and Alex Orailoglu and Cetin, {Ahmet Enis}",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/ACCESS.2019.2949740",
language = "English (US)",
volume = "7",
pages = "155701--155710",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Detecting Gas Vapor Leaks Using Uncalibrated Sensors

AU - Badawi, Diaa

AU - Ayhan, Tuba

AU - Ozev, Sule

AU - Yang, Chengmo

AU - Orailoglu, Alex

AU - Cetin, Ahmet Enis

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Chemical and infra-red sensors generate distinct responses under similar conditions because of sensor drift, noise or resolution errors. In this paper, we develop novel machine learning methods for detecting and identifying VOC and Ammonia vapor from time-series data obtained by uncalibrated chemical and infrared sensors. We process time-series sensor signals using deep neural networks (DNN). Three neural network algorithms are utilized for this purpose. Additive neural networks (termed AddNet) are based on a multiplication-devoid operator and consequently exhibit energy efficiency compared to regular neural networks. The second algorithm uses generative adversarial neural networks so as to expose the classifying neural network to more realistic data points in order to help the classifier network to deliver improved generalization. Finally, we use conventional convolutional neural networks as a baseline method. Our findings indicate that using raw time-series data obtained from uncalibrated sensors and processing them using deep-learning-based methods yield better results than using hand-crafted feature parameters.

AB - Chemical and infra-red sensors generate distinct responses under similar conditions because of sensor drift, noise or resolution errors. In this paper, we develop novel machine learning methods for detecting and identifying VOC and Ammonia vapor from time-series data obtained by uncalibrated chemical and infrared sensors. We process time-series sensor signals using deep neural networks (DNN). Three neural network algorithms are utilized for this purpose. Additive neural networks (termed AddNet) are based on a multiplication-devoid operator and consequently exhibit energy efficiency compared to regular neural networks. The second algorithm uses generative adversarial neural networks so as to expose the classifying neural network to more realistic data points in order to help the classifier network to deliver improved generalization. Finally, we use conventional convolutional neural networks as a baseline method. Our findings indicate that using raw time-series data obtained from uncalibrated sensors and processing them using deep-learning-based methods yield better results than using hand-crafted feature parameters.

KW - additive

KW - and generative adversarial (GAN) neural networks

KW - convolutional

KW - sensor drift

KW - time-series data analysis

KW - VOC gas leak detection

UR - http://www.scopus.com/inward/record.url?scp=85074730681&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074730681&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2949740

DO - 10.1109/ACCESS.2019.2949740

M3 - Article

AN - SCOPUS:85074730681

VL - 7

SP - 155701

EP - 155710

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8883150

ER -